Short-term Wind Speed Prediction Of Wind-blown Sand Field

Sandstorm has always been one of the typical environmental problems in the Northwest,and it is also one of the hot topics of concern to the whole society.The core problem of sandstorms is the sandstorm movement.The sandstorm movement itself is the transport movement of sand particles to the near-surface layer.In view of the fact that wind is a key factor in the movement of wind and sand,it has important practical and academic value for the study of wind speed characteristics and wind speed prediction.At present,commonly used wind speed prediction methods include time series method,Kalman filter method,neural network method and empirical mode decomposition method(EMD).Each method has its own advantages and disadvantages.Based on the data of wind-blown sand field observation,this paper comprehensively uses ARIMA model and EMD algorithm for numerical simulation and calculation analysis.Firstly,based on the ARIMA prediction model,the single-step time model is predicted for the downwind and crosswind wind speeds at 5m height,and compared with the measured data.The results show that the prediction results of the two wind directions are ideal.On this basis,try to use the single-step prediction model estimation instead of the actual measurement,and establish a multi-step time prediction model for wind speed.The results show that within the three-step prediction range of wind speed,the predicted value generally meets the engineering accuracy requirements,but when the forecast exceeds three At the step,the prediction accuracy is greatly reduced,and the predicted wind speed in some time periods does not match the actual.Secondly,an improved EMD wind speed prediction model is proposed,and wind speed prediction and analysis are carried out.From the wind speed prediction results of the previous ARIMA model,it is found that the predicted wind speed has a certain delay,and the prediction accuracy at some wind speed break points is not ideal.Therefore,combined with the respective advantages of EMD model and ARIMA model,an improved EMD wind speed prediction model is proposed.Specifically:(1)EMD decomposition processing of the original wind speed data,obtaining a finite number of relatively stable components,that is,reducing the mutual influence between the components,and highlighting the variation of the original data in the details;(2)According to the frequency variation characteristics of each component,it is divided into three parts: high frequency,intermediate frequency,low frequency and three components.The three components are respectively subjected to ARIMA modeling prediction.(3)The prediction results of the three components are superimposed to obtain the final prediction data.The results show that the improved EMD model has a good prediction effect,which significantly reduces the delay of the predicted wind speed and further improves the prediction accuracy of the wind speed.From the prediction results and measured results of 5.0m,7.2m,10.3m and 14.65 m at different heights,it is found that the single point data and the relationship between the wind speeds between the two points are consistent with the actual change law,and at the same time The wind speed prediction during the sandstorm was studied and the prediction effect was good.This provides a theoretical analysis and data analysis basis for the study of wind-sand movement mechanism and engineering practice.